Respiratory system mechanics assessed by Artificial Neural Networks

Artificial Neural Networks (ANNs) are suitable for different uses in the fi eld of respiratory mechanics applied to mechanical ventilation (MV), The au thors present experiments aiming at extracting lung mechanics variables fro m respiratory tracings by ANN-based methods. Some new ventilation modes aim to the best mechanical coupling between patient and ventilator, However th ese technologies are very difficult to be applied correctly if there is not a reliable source of information on the actual mechanics of the connected patient. The capability of interfacing complex variables and the robustness of their performances, suggest for the ANNs a possible use in future close d-loop MV strategies in which they could work as an integrating tool betwee n the patient mechanics and the ventilator.